Question

A healthcare company needs to label 50,000 chest X-ray images to train a disease-detection model. The images contain protected health information (PHI) and cannot be shared with anyone outside the organization. The task requires domain expertise to distinguish subtle abnormalities, and the company has a small team of in-house radiologists. Which SageMaker Ground Truth workforce option best meets these requirements?

Accepted Answer

Use a private workforce composed of the company's in-house radiologists

Answer

Use the Amazon Mechanical Turk public workforce to maximize labeling throughput at the lowest cost

Answer

Use a third-party vendor workforce from the AWS Marketplace to scale the labeling effort

Answer

Use automated data labeling (active learning) exclusively so no humans view the images